A NEW INTRAMOLECULAR REARRANGEMENT
Oct. 20, 1952
The amines obtained by reducing the oximes of benzoin, diisopropyl ketone, p-bromoacetophenone and cyclohexanone were characterized by conversion to their hydrochlo-
[CONTRIBUTION FROM
THE
COLLEGE
OF
5187
rides. a-Phenylethylamine was characterized by conveTsion to the oxalate. CHAPELHILL, NORTHCAROLINA
LIBERALARTSA N D SCIENCES, TEMPLE UNIVERSITY]
A New Intramolecular Rearrangement1 UY WILLIAM T. CALDWELL AND GEORGE C. SCHWEIKER RECEIVEDM A Y28, 1952
A new example of intramolecular rearrangement is described. It is proved t h a t hydrolysis of l-(p-nitrophenoxy)-2acetoxy-3-phthalimidopropane (VI) and l-(o-nitrophenoxy)-2-acetoxy-3-phthalimidopropane( V I I ) gives 1-(p-nitroanilinoj2,3-propanediol ( V I I I ) and 1-(o-nitroanilino)-2,3-propanediol (IX).
In the final step of an attempted synthesis of 1-
( p -nitrophenoxy) -2-hydroxy- 3 - dichloroacetamido-
propane (an isomer of Chloramphenicol) i t became apparent t h a t the expected intermediate primary amine had not been obtained in spite of excellent analyses in accord with its molecular formula. l-Phenoxy-2-hydroxy-3-chloropropane(I) was made as described by Boyd and Knowlton2 and by Levas and Lefebvre, and treated with potassium phthalimide using dimethylformamide as solvent4 to prepare l-phenoxy-2-hydroxy-3-phthalimidopropane (11). l-Phenoxy-2-hydroxy-3-aminopropane (111) was made from I according to Boyd’s directions,6 and from I1 by treatment with hydrazine, hydrochloric acid and sodium hydroxide.
I1 fuming j
NOn
A
1
OCH~CH(OAC)CH~S VI1
I
OCHzCH(0Ac)CHzS
4
NH~NH~
HCIJ
fi v
OCHzCHOHCHzNHz.HCl-+
”
VI
When the mixture of VI and VI1 was hydrolyzed in the manner used to hydrolyze 11, only a 57, yield of a mixture of 1-(fi-nitroanilino)-2,3-propanediol (VIII) and l-(o-nitroanilino)-2,3-propanediol (IX) was obtained. It was surprising to find, however, that the mixture of VI and VI1 gave good
OCH2CHOHCH2h”z I11
All attempts to prepare a crystalline diacetyl derivative of I11 failed; however, i t was found that acetylation of I1 proved an excellent way of protecting the amino and hydroxyl groups during subsequent nitration. Nitration of l-phenoxy-2acetoxy-3-phthalimidopropane (IV) with cold red fuming nitric acid gave 1-(2,4-dinitrophenoxy)-2acetoxy-3-phthalimidopropane (V), but the action of yellow fuming nitric acid on IV gave a mixture of l-(p-nitrophenoxy)-2-acetoxy-3-phthalimidopropane (VI) and 1-(o-nitrophenoxy)-2-acetoxy-3phthalimidopropane (VII) . (1) Taken from a thesis submitted by George C. Schweiker in partial fulfillment of the requirements for the degree of Master of Arts, (2) D. Boyd and H. Knowlton, J . Chcm. Soc., 96, 1802 (1909). (3) E. Levas and H. Lefebvre, Compt. rend., 119, 555 (1946). (4) J. Sheehan and W. Bolhofer, THIS JOUNAL, 71, 2786 (1950). (5) D. Boyd, J. Chcm. Soc., 97, 1791 (1910).
VI PU’HzNH2 HClI
or
,,,OH
Tu’aOH
“CHnCHOHCHzOH VI11
+
NHCH2CHOHCB2OH IX
yields of a mixture of VI11 and IX upon hydrolysis hours. The vermilion ortho compound (111.p. 7 Z - i : 3 ' ) and its yellow para isomer (m.p. IOS-1O!)") with relatively dilute sodium hydroxide solution. "111 was obtained in pure form by extraction of resemble in color and m.p. the compounds p-nitrothe ortho and para mixture; IX was obtained pure anilinoethanol (117.1). 110-1 10.5") and o-nitroanilby chromatographic adsorption. \'arious tests inoethanol (n1.p. 76-76.5" J prepared by Kremer"' and \-arious others by heating aminoethanol in the and reactions showed that hydrolysis of 1.1 and \'I1 had riot produced the expected l-(p-nitrophenoxy)- presence of anhydrous sodium carbonate with the 2-hydroxy-3-aminopropane and 1-(o-nitrophenoxy)- respective chloronitrobenzenes. IVeddige apparently could not isolate the hydro2-hydroxy-8-aminopropane. In particular, failure of methyl dichloroacetate to bring about dichloro- chloride of either X or X I and he stated that the acetylation forced US to consider the possibility presumed hydrochloride of XI in a concentrated of an intraniolecular rearrangement t o a compound hydrochloric acid solution was decomposed by no longer containing a free primary amino group. the addition of water. Furthermore, when he warmed X I with a mole of benzoyl chloride two This possibility became a probability when search of the literature revealed that the m p. substances were formed; one compound was a of IX was close to that of l-jo-nitroanilirictj-2,:)- iiionobenzoyl derivative, and the other was a dipropanediol reported by Karrer, Saloinan, Schopp benzoyl derivative which in MTeddige's opinion was arid Schlitter.6 VI11 and IX were made accord- the N,N-dibenzoyl derivative of XI. I t does not ingly by the method of Karrer, et d., aiid the coi11- seem likely that X I would form a K,N-dibenzoyl pounds prepared in these different wdys were coni- derivative under these conditions; on the other pared. llixed m.p. determinations, infrared ab- hand, if XI had rearranged, in a fashion analogous sorption spectrd and X-ray diffraction patterns, as to the rearrangement proved in this paper, to fonn well as melting points, solubilities and colors showed o-nitroanilinoethanol, it is easy to understand how a that the compounds prepared by these different monobenzoyl derivative and a N,O-dibenzoyl derivative could be formed. Also, if X and X I procedures were identical. This rearrangement is a new example of intra- had rearranged to form the corresponding nitromolecular nucleophilic displacements, inany of anilino compounds the instability of the hydrowhich are known collectively as the Smiles re- chloride salts would then be explained. arrangement'; it appears to be the first recognition I\*e wish to thank Eli Lilly and Company for of a Case of an amino group on an alkyl side chain data on infrared absorption and X-ray diffraction, displacing the alkoxy group from an aromatic and the Temple Cniversity Committee on Research nucleus activated for nucleophilic attack by the and Publications for a Grant-iri-Aid. presence of a nitro group. Evans and Smiles,' Experimental" and Bunnett and Zahler,h have summarized the several sorts of Smiles rearrangements which have l-Phenoxy-2-hydroxy-3-phthalimidopropane(II).-Sheebeen observed in the following way where in X- hati and Bolhofcr' reported t h a t dimethylformamide is a n C C 1-H the group I-H is separated from X by two excellent solvent for the Gabriel reaction. Accordingly, I1 made from l-ph~noxy-2-hydroxy-3-chloropropane~~~ intervening carbon atoms. I n previouslv des- i\vas ~ n t lpotassiuiii phthalimitie using dimethylformamide in a
11
cribed rearrangements, Y H has stood for KHacyl, etc., and X for SOs, S, 0, etc.* In the present rearrangement, Y H is SHa(alky1i and X is 0 and the intervening chain contains three atoms of carbon. I n the light of this rearrangement, it is logical to assume t h a t rearrangements of other nitrophenoxy alkyl amines may have escaped detection. For example, this would seem to be true of the cotnpounds l-(p-nitrophenoxy)-2-aniinoethane (X) and 1 -!o-Iiitrophenoxy)-"-aminoethane (XI ) reported by M-eddige.!' O ~ Z ~ - ~__~ , O C H ? C I I . T H . : ~
s SI In the preparation of X and XI, Weddige heated
the corresponding nitrophenoxy- P-bromoethanes with alcoholic ammonia a t 100-120" for several (ti) 1'. Karrer, IT. Saloman, I;. Scliopii and Is Schlitter. I I f l : ' . ('hit?!, ..IL/,I, 17, 1163 (1034). ( 7 ) W. l l r a i i h a n d S . Smilc~s.J ( i i e t i i S O L , 181 l l W . i l
manner analogous t o that described by Sheehan and Bolhofer. Po\r-dery whitc I1 \viis obtained in 76% yield, m.p. 108-110'. .liiol. Calcd. for C I ~ H I S O ~ SC:, 68.67; H, 5.09; S , 4.71. Found: C, W.77; 13, 3.15; S ,4.98. I l-Phenoxy-2-hydroxy-3-aminopropane(111) was prepared by hydrolyzing I 1 i i i :i manner similar t o t h a t used by Sheeh a n aiid Bolhofer' t o hydrolyze various other phthalimido compounds; also, I I 1 was prepared from l-phenoxy-2IiJ'drosy-3-chloroprop:tne2." described hy Boyd,5 which ble hydrochloride of 111. A required the isolation of the mixed m . p . of the compounds prepared in these two differcnt ways resulted in no depression; m . p . 96-97". l-Phenoxy-2-acetoxy-3-phthalimidopropane (IV) .--The acetyl derivative of I1 was prepared in the usual manner using acetic anhydride and anhydrous pyridine. K h i t e , crystalline I\' was recovercd in 92.5% yield, m.p. 85-87'. A n a l . Caictl. for ClgHl;OjS: S, 4.13. Found: S , 4.30. 1-(2,4-Dinitrophenoxy)-2-acetoxy-3-phthalimidopropane ( V )rcsulted from thc addition of 11: t o red fuming nitric acid (sp. gr. l . 6 j kept a t a temperature of - 3 to 3'. After 1 0 iiiinutcs a t this temperature the solution was poured into ice-water and the product filtered off and washed with 5% sodiuni hicarbonatc and water. Recrystallizations from alcohol, chloroform, ethyl acetate-chloroform and benzene gavc ii 2;3.770 yield of pale yellow crystalline T', m.p. 191196". A n a l . Calcd. for C1qHi305S3:N, 9.79. Found: N, 9.74. Nitration of IV using yellow fuming nitric acid gave a mixturc of 1-( /,-nitropheriosy'i-~--;lcctol;y-3-phthalimidopro]~allc ~
' Ill1
~.~
.
C. K r e m c r , ' l r i i i i I < ) K( S A L , 61, 1321 (I9:W).
Oct. 20, 193.2
~-PROPYL-~-ETHYL-~-NITROSOOXAZOLIDINE FROM %AMINO-P BUTANOL
(\:I ) and I-( o-nitrophenoxy)-2-acetoxy-3-phthalimidopropane ( V I I ) . A solution of yellow fuming nitric acid (sp. gr. 1.5) in glacial acetic acid was added t o a solution of I\in 4 parts of glacial acetic acid to 1 part of acetic anhydride a t 80°, and the resulting red solution was kept a t 80' for another hour before it was poured into ice-watei. Filtration and washing with 5% sodium bicarbonate and water, followed by trituration with boiling ethanol produced a crystalline mixture of \'I and VI1 which was recrystallized from benzene. The pale yellow mixture was obtained in 81.6% yield, m.p. 140-146'. Separation of the ortho and para isomers a t this stage seemed inadvisable because of the great loss of material during recrystallizations of an analytical sample. The rzsults of a nitrogen analysis on this sample, m.p. 158-160 , were in keeping with the composition of either \'I or VI1 or a mixture of the two. Anal. Calcd. for CigH1607N2:S , 7.29. Found: N , 7.18. I-(p-Nitroanilino)-2,3-propanediol(VIII). A.-The mixture of V I and VI1 (m.p. 140-146') was hydrolyzed in the mannei4 used t o hydrolyze 11, except t h a t the refluxing time with hydrazine ( 2 moles in this case) was lengthened to 6 hours, and refluxing with 1 to I hydrochloric acid was lengthened to 8 hours. The dark oil so produced yielded only 5.1Y0 of orange-yellow crystalline material, m . p . 112114', after many extractions and fractional crystallizations with various solvents. .4naZ. Calcd. for C y H i u 0 4 S ? :C, 50.94; H, 5.70; S, 13.20. Fpund: C, 50.74; H , .5.70; S ,13.15. I t was found later that this constant melting material was still a mixture of the ortho and para isomers. B.-Sixty grams (0.156 mole) of the mixture of V I and 1-11(m.p. 110-146') was mixed with 20 g. of sodium hydroxide (0.50 mole) in 160 ml. of water and refluxed for 1 hour. As reflux temperature was approached, the two phase system changed into a deep red solution after approximately 5 minutes of heating. After cooling, orange crystals were filtered off and recrystallized from water and ethyl acetate. (Acidification of the filtrate produced phthalic acid.) The product was extracted with three small portions of chloroform, and the material which did not dissolve was extracted
with an alcohol-chloroform mixture.
3189
The yellow crystals
(\:III) still remaining undissolved were recrystallized from alcohol-acetone and alcohol-chloroform; m.p. 126-127". From many kactional crystallizations of the chloroform extracts, two separate piles of material were collected and recrystallized separately from ethyi acetate. The yield of yellow VI11 was 2 g., m.p. 126-127 ; the yield of the orangeyellow mixture of VI11 and its ortho isomer ( I X ) was 18 g., m.p. 106-112'. C.-\'I11 was made in 6% yield, m.p. 126-127", in a fashion analogous to t h a t described by Karrer,G et al., for the preparation of its ortho isomer ( I X ) , using 1-amino-2,3propanediolI2 and p-riitrochlorobenzene. Anal. Calcd. for C3H1204L-2: N, 13.20. Found: S , 13.06. A mixed m.p. determination of VI11 prepared as described in B and C gave no depression. l-(o-Nitroanilino)-2,3-propanediol(IX). A.-Recrystallizations of the mixture of VI11 and IX, as prepared in B above (m.p. 106-112O), from chloroform gave a constant melting mixture, m.p. 112-114°, and the yield decreased t o 1 g. However, orange-yellow IX, m.p. 117-118°, was isolated from this constant melting mixture by chromatographic adsorption. The mixture was deposited upon alumina from chloroform solutions, developed with 10% absolute ethanol in chloroform, and eluted .with methanol. B.-IX was made in 11% yield as described by Karrer,fl el a/.,m.p. 117-118'. A mixed m . p . determiuation of IX prepared as described in A and B gave no depression. I-(o-Nitroanilino)-2,3-propanediolhydrochloride could not be isolated from a n aqueous hydrochloric acid solution of IX by evaporation of the solvent in a vacuum desiccator. However, the white salt was obtained by passing hydrogen chloride gas through an ether solution of IX, containing a small amount of methanol. The salt turned pale yellow during filtering and washing with ether, and a portion in moist air very quickly decomposed into the original orangeyellow IX. (12) L. and E. Knorr, C h e m Brv , 32, 730 (1809).
PHILADELPHIA 22, PESNA.
[COSTRIBUTION FROM
THE
LIETCALF CHEMICAL LABORATORIES, BROWNUXIVERSITY ]
2-Propyl-4-ethyl-3-nitrosooxazolidine,a Novel Product from the Nitrous Acid Deamination of 2-Amino-1-butanol BY HAROLD R. NACEAND MORTON H. GOLLIS' RECEIVEDM A Y22, 1952 The nitrous acid deamination of 2-amino-1-butanol in acetic acid yields 2-propyl-4-ethyl-3-~~itroso6xazoliditie (49-52%), 1,2-butanediol (470) and crotyl alcohol (4%). When the reaction is carried out in hydrochloric acid, 2-chloro-I-butanol (2%) is also obtained. The nitrosooxazolidine is cleaved by acid t o n-butyraldehyde and the starting amino alcohol. The nitrosooxazolidine was svnthesized b v nitrosation of the oxazolidine, formed by the condensation of n-butyraldehyde with 2amino-1-butanol.
Many examples of the reaction of p-amino alcohols with nitrous acid have been reported.2 I n the majority of the cases studied products with rearranged carbon skeletons have resulted, the products being analogous to those of the pinacol rearrangement. However, the reaction has been complicated by one or more factors: (a) The presence of one or more aryl groups on the substituentbearing carbon atoms has superimposed migratory aptitude and steric effects on the course of the rearrangemenL3 (b) The substituents were at(1) Research Corporation Fellow, 1951-1952. (2) G . W.Wheland, "Advanced Organic Chemistry," John Wiley and Sons, Inc , New York, N. Y.,1949, Chap. 12. (3) For further discussion and references see P. I. Pollak and D. Y . Curtin, THISJOURNAL, 72, 961 (1950).
tached to carbon atoms incorporated in alicyclic structures, thus bringing into play the effect of polar and equatorial bonds on the ease and direction of the rearrangement.4 Simple aliphatic amino alcohols would be free of these complicating factors, and a study of them should facilitate the understanding of the more involved cases. A search of the literature revealed only a few examples of the reaction of simple aliphatic amino alcohols with nitrous acid. Krassusky and Dudas obtained in unstated yields pinacol and pinacolone from the reaction of 2,3-dimethyl-3-amino-2-b~tanol with nitrous acid. Neuberg and Rewald6 (4) G. E. McCasland, i b i d . , 78, 2293 (1951). (5) K. Krasnusky and L. Duda, J . p r a k l . Chcnt., 121 77, Y6 (leos). ( 6 ) C . Neuberg and R . Rewald, Biochem. Z.,67, 132 (1914).
